The problem of the study was to create a valid and reliable instrument which would measure the diabetes knowledge of college students.To reach this goal, a pool of questions containing 49 items was developed after an extensive review of the literature which addressed diabetes risk factors, diabetes knowledge, college students and health knowledge, instrument development and knowledge measurement. A nine member jury of experts composed of health education professors, those knowledgeable in instrument development, and diabetes health care providers determined content validity for the pilot test instrument which reduced the pool to 34 items. The instrument was then pilot tested on 46 Ball State University undergraduates in a personal health course. The results were analyzed and had a whole test reliability, using the Kuder-Richardson 20, of .76.General education students (n = 522) at Central Michigan and Ball State Universities were the subjects for the administration of the revised instrument. The revised instrument included 27 items. The final instrument contained 26 items and had a whole test reliability coefficient, using the Kuder-Richardson 20, of .78. In addition, it discriminated between the upper and lower 27% groups in terms of total test score, had item-test reliability coefficients of .11 or greater, and the item difficulty ranged from .17 to .87. / Department of Physiology and Health Science